Fetching the paper…

SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion · Around